The MOSAIC Land Surface Model of Koster and Suarez has been coupled to the Goddard EOS(GEOS) GCM, which is a component of the GEOS Data Assimilation System. The behavior of the coupled system in simulations (GCM alone) and assimilations is examined and compared to available conventional and satellite observational climate estimates to understand its impact on the GCM and DAS mean climate states. The direct impact is seen in the decreased rates of latent heat flux and increased rates of sensible heat flux, which is a result of the increased resistence to moisture flow though the vegetation stand described by many Soil-Vegetation- Atmosphere-Transfer (SVAT) models such as MOSAIC. The linear response to the drier and warmer near surface condition is a decreased continental precipitation and an increased ground temperature. The robustness of the direct impact of the LSM and the linear response is seen in both GCM and DAS atmospheric climates. The indirect response, however, is different in the GCM and the DAS, and is related to the ability of the LSM and surface layer parameterizations to properly absorb and retain the influence of the data brought in by the GEOS Physical-Space Statistical Analysis System (PSAS) during the assimilation